beatabix wrote:i read a paper recently that seemed to suggest that limiting rotation to a single axis was relatively easy, something like just setting one strip in the array to 0,0,0,0 or something like that.
it went on to say that putting angle limits on a joint wasn't a particularly great leap either, though more complicated than that.
this was using a jacobian matrix, like blender, i think.
anyway, the paper said it's not too difficult, compared to some of the other stuff you might want to do with a kinematic system.
This is an oversimplification unfortunatley.
While deciding whether or not a rotation is outside the space defined by joint limits is relativley simple there are a host of other issues that need to be taken into account. For instance what about interpolation of the rotations such that over time no rotation state violates the joint parameters? Also not too difficult in theory but add into this the constraint of making it interpolate without pops or snaps and
Theres all kinds of other things too, but well you get the idea.